This PDF is a selection from an out-of-print volume from the National Bureau of Economic Research

Transcription

1 This PDF is a selection from an out-of-print volume from the National Bureau of Economic Research Volume Title: Risk Elements in Consumer Instalment Financing Volume Author/Editor: David Durand Volume Publisher: NBER Volume ISBN: Volume URL: Publication Date: 1941 Chapter Title: Scope and Purpose of the Study Chapter Author: David Durand Chapter URL: Chapter pages in book: (p. 9-21)

2 1 Scope and Purpose of the Study SotE form of credit investigation is essential to any loan transaction, whether large or small, personal or business. When making an investigation, a lender does not attempt to predict the applicant's repayment record precisely, hut he does undertake to distinguish the applicants who are likely to repay from those who are not. In this attempt, he obtains pertinent information on such matters as borrower's character and reputation, borrower's financial condition, intended use of the funds, and type and value of any collateral that may be offered as security. From an analysis of this information, the lender evaluates the proposed transaction; in doing this, he relies on his experience and intuition, and in addition he may weigh the available facts according to some mechanical routine. In the present study, we are presenting a statistical approach to the analysis of certain pertinent risk elements in consumer instalment financing. The purpose of the study is twofold: first, to analyze the significance of some credit factors generally considered important by credit men, to discern which of these factors have been associated in the past with bad loans, and to determine whether or not this information can be used to predict the course of future transactions; second, to experiment with the use of statistical methods that may be applied to the problem of credit risk selection. Statistical methods, particularly small sample techniques. have not heretofore been used to any great extent in credit analysis, and their value in such analysis is not well known. This is regrettable, for the orderly recording of experience, which 9

3 lo RISK IN INSTALMENT FINANCING is the esseiace of the statistical nicthod, should titl Coflsjder ably in solving risk problems. The statistical method evolved in this study has been designed to permit analysis of loan experience without undue expenditure of labor, and its application is simple enougi, we believe, to be of practical value to credit executis' in solving some of their own particular problems. But the method has its shortcomings as well as its advantages. Al. though it may be effectively used for testing various general hypotheses and for discovering unsuspected relations be. tween credit information and risk experience, it is an inferior method for determining what type of credit policy will pro. duce lowest costs and highest profits. Detailed consideration of the proper approach to the questions of costs and credit policy will be given in a later chapter. CHARACTERISTICS OF CONSUMER INSTALMENT CREDIT Provision for periodic repayment is the most general characteristic of consumer instalment credit, but a number of others are fairly typical. Loans are for the most part small; nearly all are substantially less than a thousand dollars in size, and the majority are less than five hundred dollars. The proceeds are usually employed by ultimate consumers to purchase goods, to finance a current deficit, or to meet a personal emergency; but they are also used for business purposes.' The security for these loans does not ordinarily consist of stocks and bonds or mortgageable real estate, which would be acceptable collateral for a loan at almost any commercial bank. The usual forms are endorsements by comakers, con- 'Loans used For consumpijon purpae cannot always be distinguished easily from those used for business purpo. When a small business enterprise buys a truck on the instalment plan, a business loan is undeniably involved. On the other hand, while the purchase of a passenger automobile would usually be considered a consumer loan, many passenger cars are used partly for business purposes.

4 SCOPE AND PURPOSE OF THE STUDY ditional sales contracts, and chattel mortgages on automobiles or household goods; in many cases there is no security at all. Interest rates on consumer loans are on the whole higher than on business loans, but the costs of doing business are also higher. A number of different types of institutions make consumer instalment loans; the main ones are industrial banking companies, commercial banks, personal finance companies, sales finance companies, retail dealers, and credit unions. In general, consumer instalment credit transactions may be classified in two broad groupssales finance transactions and personal finance transactions. The former are handled largely by sales finance companies or retail merchants, and to a lesser extent by commercial banks and industrial banking companies; cash loan transactions are usually handled by personal finance companies, commercial banks, industrial banking companies, credit unions, and a number of lesser agencies. In sales finance, some retail merchants handle the entire transaction theinselvcsthe investigation of risk, the arrangement of terms, and the collection of payments; others investigate the risk, arrange the terms, and then dispose of the customer's obligations to one of several agencies that discount and collect such paper; and still others leave all the arrangements to their financing agencies. In personal finance, the borrower applies to one of the available lending agencies for a loan of cash, and this agency, if it approves the application, assumes the task of collecting the account. The borrower uses the loan for some specified purpose or purposes; he may use it to purchase goods, in which case it may be a substitute for a sales finance transaction, or he may use it to meet emergencies or to settle pre-existing obligations. In legal discussions it is extremely important to distinguish sales finance transactions from personal finance transactions, for a debt arising out of a bona fide sale of goods on the deferred payment plan is not legally a loan and is not subject to the maximum interest statutes. For purposes of risk analy-

5 RISK IN INSTALMENT FINANCING sis however, the (listinctiofl is of little use. Although we siwil occasionally find it convenient to consider sales credit Sep. arately from personal credit, we shall usually l)refer to consider both together; we shall use the term lot-ui to include both cash loans and instalment sales transactions, and the term lender to include sales finance companies and retail merchants as well as cash lending agencies. On the whole, consumer instalment lending iflstjtutjoi tend to establish themselves in urban areas and to grant credit to urban and suburban dwellers with moderate and regular incomes. These facts are reflected in the vocational coluposi. tion and income distribution of their clientele. Persons of all occupations are included, from unskilled laborers to the higher-paid business and professional classes. The number of farmers included, however, is relatively small in comparison with the general population.2 Farmers tend to seek other forms of credit since on the whole they have seasonal incomes unsuited to meeting regular monthly payments; as entrepreneurs, they can usually satisfy their personal, as well as their business needs with non-instaliiieiit loans from commercial banks and from a number of other institutions specially set up to provide credit to farmers. The highest and lowest income classes also are tmderrepresented. Although the income distribution covers a wide range, substantially more than half of all the cases in the samples on which this study is based represent borrowers with incomes between $1200 and $3000; and among purchasers of used cars and borrowers from personal finance companies. more than 75 percent of each group falls within this income range.8 The proportion of the total borrowing public included in this income group is considerably higher than the analogous proportion of the country's total population; the 2 No farmer group is given in the occupational distribution presented in Table 13. pp , because the number 01 farmers was too small to justify a separate classification. 3See Table 4, pp

6 SCOPE AND PURPOSE OF THE STUDY 13 latter proportion is oniy about 47 percent.' The well-to-do are underrepresented because they have relatively less need for consumer loans than do other classes; when they do require financing, they can usually obtain lower interest rates on non-instalment bank loans either unsecured or secured by stocks and bonds or real estate; and in any event they rarely have to resort to the higher-rate loans secured by comakers or chattel mortgages. The very poorest classes, the unemployed and the indigent and those with incomes substantially less than $1200, are underrepresented among borrowers by necessity rather than by choice; they are excluded from credit facilities by lenders' credit standards. Instalment loans are intended to be repaid out of income, and the main concern of lenders is that their borrowers shall have a means of livelihood. Therefore, customers usually must be employed or engaged in business, and they must receive a remuneration sufficient for their support. Probably every lender operates on a set of minimum income standards. These standards are precisely formulated, and are capable of being waived in special circumstances; nevertheless, they serve to prevent paupers and near paupers from availing themselves of loan facilities. This percentage, which was estimated for by the National Resources Committee, refers to the income class of $1000 to $3000. It includes all single persons and non-relief families. See Consumer Incomes in the United States (1938) Table 8, P. 25. The picture we have presented is based on the operations of sales finance companies, personal finance companies, commercial banks, and industrial banking companies. If we had had data for other lending agencies, particularly retail merchants, we might have shown that even lower-income classes make extensive use of instalment credit. Retail dealers extend credit to very low-income classes, particularly in communities where dothing, cheap radios. and other low-priced goods are sold on the instalment plan. Furthermore, mail-order houses extend considerable amounts of instalment credit to farmers. The limits of the interval $1200 to $3000 are not particularly significant since we do not know the precise points at which the overrepresented middle group ends and the underrepresented extreme groups begin. However, the chosen limits are convenient for broad illustrative purposes. The records of most lenders show a few cases of borrowers with little or no income, but these cases can usually be explained in terms of family or

7 14 RISK in INSTALMENT FINANCIN RISK SELECTION Because consumer credit is designed to be available to the general masses, and because small SLImS arc involved the detailed and exhaustive investigation of credit standing characteristic of most commercial loan transactions is not typical of consumer instalment credit transactions; a more routine and less individual procedure is necessai-y, The lender has to content himself with a rather limited amount of credit information. In essence, the process of risk selection is more one of rejecting a small nunther of undesjrables from a large body of generally satisfactory applicants than one of selecting a limited number of high-grade risks. Despite this fact, the consumer credit business has grown rapidly, and consumer lenders usually succeed in avoiding undue losses. Risk selection is effected not only by the interviewing of applicants but also by other methods. For instance, after deciding what class of customers it wishes to deal with, a lending firm can direct its advertising to that class; and by making known its general requirements for loans, it can dissuade a great number of unsatisfactory risks from applying for credit. Standards of acceptability doubtless vary considerably from lender to lender, but certain broad principles appear fairly universal. The minimum income restriction, mentioned above, seems to be one of the fundamentals of risk selection. Another fundamental is time lender's insistence upon obtaining borrowers who are reasonably honest and fairly responsible in meeting their obligations To assure this result, a good share of the credit investigation consists in checking other personal reiauonshj. Examples signature of her husband, are a wife who borrows on the to- a sort who succeeds in buying because of his father's a car on credit credit standing, his expenses while or an unemployed person who meets seeking employment by co-signatures with the proceeds of a loan secured of friends. These they do occur the cases do not occur frequently, and when reported income is misleading.

8 SCOPE AND PURPOSE OF THE STUDY 15 applications for false statements, consulting credit bureaus and previous creditors for signs of irresponsibility, and searching court records. Apparently some lenders reject a credit applicant if he has ever failed to meet financial obligations or if he falsifies his application; others, however, may l)e much more lenient. In addition to minimum incomes and satisfactory past payment records, applicants are expected to possess other qualifications, and to supply on their applications information concerning these qualifications. This information, which concerns age, marital status, number of dependents, stability of residence and occupation, assets and liabilities, etc., is indicative of what credit officials generally consider important. The items collected for our analysis of risk elements are reasonably representative except for one item; no information is available on the important matter of past payment record. Although a study of application forms demonstrates the items of information required, it does not even suggest the relative importance that credit officials attach to these items. Consequently, two sets of questionnaires were sent to credit officials, to determine how operating lenders evaluate credit factors. One set went to commercial bankers operating personal loan departments, and the other to retailers in several lines of business; 126 replies were received from the former and 688 replies from the latter. The bankers, in their questionnaires were requested to state what credit factors they considered important and to list them in order of importance. No hints were given concerning the answers expected except that reference to income was not to be included. Although the replies showed amazing variation, many of the differences seemed to be merely in terminology. The answers were classified into 15 convenient categories, such as character and reputation stability of employment, borrower's liabilities, etc. These categories were then given ratings on a point system; a score of 5 was given

9 RISK IN INSTALMENT FINANCING each category every time it was listed first in importanc by a banker; a score of 4 was given every time it was listed second; 3 was given for third place; 2 for fourth place; and I for fifth place. These 15 categories were then consolidateti into 5 broad groups; for example, the similar items of past paymett record, character and reputation, and credit rating were combined to form the broad group, moral characteristics A score for each broad group was then determined by computing a simple sum of the scores of the component items. All the scores, both single and combined, were then con'erte(j to index numbers, with the total score of the vocational characteristics as the base figure (i.e., 100). The results are presented in Table I. As the table stands, moral characteristics are most highly emphasized, financial characteristics al-c next, and vocational characteristics are a very close third. But this presentation of results may be extremely misleading, for the table could be rearranged to show a very different order of importance. The table was al-ranged in its present form because of certain resemblances among the grouped categories. Character and reputation, past payment record, and credit rating. all seemed to us to be virtually identical, and hence they were grouped together; but they may not have seemed identical to the bankers who filled in the questionnaires. Thus, although the present form of the table shows that moral characteristics are most important, with financial characteristics next, and vocational characteristics third, other interj)rctatiofls are possible; for example, stability of occupation may he considered most important, with past payment mecoi-d second, acter and reputation and char- third. The questionnaire for retail merchants was arranged differently. Six Specific credit factors were listed on the questionnaire, and the merchants Were requested to indicate the relative order of importance of these were also requested factors. The merchants to list any other factors that they con-

10 SCOPE AND PURPOSE OF THE STUDY 17 TABLE I INDEx or REwivE IMPoRrcE ATrACHED To Vious CREDIT FAcrroas OTHER T1w4 INCOME BY 126 CoMMERCIAL BKs' Credit Factor lndexb Vocational characteristics Work performed Industry and employer Stability Financial characteristics 21 Liabilities 46 Income balance 36 Moral characteristics Past payment record 63 Character and reputation 59 Credit rating 40 Personal characteristics Age 5 Marital status 4 Number of dependents 18 Stability of residence 3 Loan characteristics Security Duration Based on a questiounaitc aurve made by the National Bureau of Economic Research. A more detailed table, showing number of times each factor was given first or second or third place, is given in National Bureau of Economic Research (Financial Research Program), Commercial Banks and Consumer Instalment Credit, by John M. Chapman and Associates (19-10) p For explanation see text, pp

11 i8 RISK IN INSTALMENT FINANCING sidered important, but few of them availed themsc1'es of the opportunity. The replies were rated by the same point system used for the bank questionnaire; and the scores were converted to relatives, with occupation and peliflafldfice of employment equaling 100. The results are shown in Table 2. In this table, item (1), occupation and permanel'ice of employment, and item (2), past payment record, appear clearly as the two most important factors, with all others distinctly secondary. But items (3) and (4) could be combined to form a group similar to financial characteristics in Table 1, and then the combined score of 90 would bring this group TABLE 2 INDEX OF RELATiVE IMPORTANCE ATTACHED TO VAR!- ous CREDIT FACTORS imy 688 RETAIL ESTABLISHMEN.I Credit Factor (1) Occupation and permanence of employment (2) Past payment record 1ndes (3) Terms conveniently adjusted to customer's income 48 (4) Additional instalment obligations to other stores 42 (5) Terms that will secure largest down payment and fastest liquidation possible (6) Length of loan contract Based on a questionnaire survey conducted by the National Retail Credit Association in cooperation with the National Bureau of Economic Research (Financial Research Program), July and August, Answers were received from 688 retail establishments situated in 157 cities in the United States and Canada. The answers were graded by giving 5 points to the credit factors ranked first in importance by the retail merchants, 4 to those ranked second, 3 to those ranked third, 2 to those ranked fourth, and 1 to those ranked fifth; the final scores were then converted to relatives with item (1) equal to 100. A more detailed table, showing replies by various t)pes of stores, is to be found in National Bureau of Economic Research (Financial Research Program), Sales Finance Corn /ajnies and Their Credit Practices, by Wilbur C. Plummer and Ralph A. Young (1940) p b For explanation, see text abose,

12 SCOPE AND PURPOSE OF THE STUDY 19 into a position of importance with the first two items; similarly, the combining of groups (3), (5), and (6) would result in a total score of 97. The two questionnaires together substantiate fairly well our views of the two fundamental requirements of risk selection. Lenders attach considerable importance to past payment record, which was one of our fundamentals. Although the questionnaires did not request information concerning minimum income, our other fundamental, its miportance may be gauged in other ways. For all practical purposes the minimum income requirement means that borrowers must be employed and self-supporting; and since the questionnaires indicate emphatically that borrowers are expected to be employed in stable positions, the matter of minimum income adequate for self-support can almost be taken for granted. In addition to employment stability and repayment record, other factors such as borrower liabilities and the relation of income to obligation are important, but no attempt is made in this study to determine the relative importance placed upon these factors by credit executives. NATURE OF PROBLEM The specific problem undertaken in this study is the analysis, by use of statistical methods, of sonic credit factors generally considered important in consumer instalment financing. The data on which the study is based were transcribed from actual loan applications, and were contributed by 37 lending concerns operating in various phases of consumer instalment credit. About 7200 loans were represented, of which about 2700 cases were supplied by commercial banks operating personal loan departments, 1400 by personal finance companies, 1300 by industrial banking companies, 1400 by automobile finance companies, and 400 by appliance finance companies. The data obtained covered a number of bor-

13

14 SCOPE AND PURPOSE OF THE STUDY 21 manly interested in reducing losses by further raising of credit restrictions. Even among the selected, high-grade risks, different degrees of excellence exist. On the basis of the findings presented here, or on the basis of a similar study of his own past loan experience, a lender may determine the characteristics of the most desirable of his present customers; he may then supplement his existing credit standards with others to obtain an even more highly selected group of customers. On the other hand, if a lender is interested in increasing the volume of loans as far as possible, the value of this studyor of any study based on past loan experienceis distinctly limited. The lender's problem is to determine which of the borrowers hitherto excluded can be safely included in the acceptable class. One possible procedure is to make experimental loans, which amounts to a temporary lowering of standards, with a possible increase in losses, and a subsequent adjustment of standards on the basis of the newly gained experience. Another possible procedure is to follow the experience of some other lender who has already relaxed credit standards. Of course, the methods of analysis used in the present study are appropriate for evaluating experience with experimental loans or for evaluating the experience of another lender.